Data source: http://housing.gatech.edu/rooms/FreeRooms.json in April 2020, showing search process for 2020-2021 year
I'm doing web-scraping this year (Fall21-Spring22 search), see this GitHub repo
I won't be running visualizations until afterwards though.
- Web scraping didn't start until after room-stay selection, and started in the middle of 2nd Year North-Ave self-selection
- North Ave data is therefore incomplete
- Web scraping data might have halted for a bit at some point, no promises. But the timestamps are all there
- Stopped web scraping after the first few days of room selection because I got bored
Q: Is my Room Selection Number high enough to get the room(s) I want?
A: IDK the answer to that. I made a document last year where ppl can put their Room Selection Number and Timeslot: https://docs.google.com/spreadsheets/d/10IjfwrKaCz9sJFydztzRTWKretjQCbet1AT58FXwBes/edit?usp=sharing, but there aren't enough data points yet. Put yours in a comment and we can figure it out, it has to be piecewise linear :)
Also, keep in mind that this is only partial data for one (pretty strange) year, so past data can't really indicate how likely you are to get a spot or not. Even GT Housing has already made some changes, like reserving only a few floors in North Ave buildings for 2nd years instead of entire buildings (probably because the data show that 2nd years didn't fully fill those spots)
Additionally, this visualization does not show restrictions on certain rooms for LLCs, undergrad/grad, etc
Q: Wouldn't this visualization be better as a webpage than some python Tkinter crap?
A: Yes, go for it. I suggest looking at using the compressed data; you don't actually need to pickle it, you could just save the timestamps as strings/ints and keep it all in json format
Q: What about data for this year? Could you update it in real-time?
A: I mean, probably. I'm sure there's a way to setup a server to log the data and push it to a github repo automatically, and have a script to compress the data so a webpage could use it. That would be an interesting project, but there are like 2 weeks before room selection and I'm kinda busy.
The data is already generated, so just cd src/ and run:
python visual.py
or
python visual2.py
- src/web_scraper/getjson.sh is a bash script I used to get the data every 10 minutes
- src/data/ holds all the data I recorded
- /src/file_time_mapper.py maps the web-scrape filenames to timestamps
- /src/generate_building_room_map.py makes a mapping of all the buildings and rooms that appeared throughout the web-scraping: /src/mappings/building_room_map.json
- /src/visual.py naively goes through all the web-scraped jsons, maps them into /src/mappings/building_room_map.json, and prints the output on a Tkinter canvas
- /src/json_parser.py and /src/test.py compress all the web-scraped data by generating an "UpdateLog" of when each room appears and disappears in the web-scraped jsons. The output is put into a python pickle file, /src/mappings/big_dict.pickle
- /src/visual2.py is similar to /src/visual.py, except it uses the compressed data /src/mappings/big_dict.pickle and has some visual changes